ERRATA. Page 12, lines 16 and 17, for one hundred read three hundred and for one thousand read six hundred. Page 17, line 2, dele first letter in the line. Page 168, line 12, page 177, lines 13 and 14, and page 271, line 10, for Lemna tnsulca read Spirodela polyrMza. Page 209, line 2 of foot-note, after hut insert represents. Page 2;'56, line 7, and page 266, line 19: snoioi n. s. has been shown to be hieroghjijhka, (^. Page 257, insert as line 8 as follows: -ken to the office produced young in ten days. The Page 272, line 13, for P. higuUatus read Pompiliis higuttatus. Page 278, Plate Y., 16, after view insert as follows: a, mentum; 6, labial rudiment; c, maxillary palpi; d, maxilla; e, labrum; /, antenna; g, eye; h, mandible. Page 286, line 11, drop initial the one line. Page 386, line 1, for Comstocki read Comstock. Page 399, line 17, for specimens read specimen. Page 411, line 10, for Michaelson read Michaelsen. Page 441, line 3 from bottom, for 66 read 68. Page 445, line 10 from bottom, for 57 read 5S. Page 466, line 1 from bottom, for Cyima read Oypris. Article VI.— On the Entomology of the lUinou River and Adjacent Waters. Fikst Paper, By C. A. Hart. INTRODUCTORY. This paper gives a part of the results of our observa- vation and study of the insect fauna of the Illinois River and adjoining waters in the neighborhood of the Uni- versity of Illinois Biological Experiment Station, at Havana, Illinois, during the first year of the Station work, as a preparation for further and more detailed observations in the same field. In order to make the account more complete and useful to Illinois students, and to give a general view of the relations of the species gtudied to the aquatic fauna of the State as a whole, the data concerning these forms afforded by the note boxes and general collections of the State Laboratory of Natural History are also here included. Distinctive characters, when known to me, are tabu- lated in the form of keys. These are necessarily imper- fect because of the small number of immature forms now known, the diflSculty of determining the true rank of the differences observed, and the frequent impossibility of harmonizing the grouping with the present classifica- tion of the imagos. I hope, however, that they may prove useful, and suggestive in further studies. It is evident that a knowledge of all stages is necessary to the formation of the best natural classification. Special thanks are due to Prof. S. A. Forbes, who originated the Station and has planned and directed its work; and to Messrs. C. H. Fernald, W. H. Ashmead, and W. A. Snow for determinations. I am also under obliga- tionB to my fellow-workeis at the Station for their kindly cooperation, and to Miss Lj'dia Moore Hart for the accurate drawings herewith presented. 150 Illinois State Laboratory of Natural History. GROUPS TREATED. The insects treated in the present paper belong to three groups: the aquatic caterpillars (Lepidoptera), the hymenopterous enemies or parasites of water insects, (Hymenoptera,) and the aquatic flies (Diptera) of several faunlies, those including the larger forms of the Ortho- rhapha, or "straight-seamed flies;" especially the crane-flie& (Tipulidae, PI. V-IX) the soldier flies (Stratiomyiidae, PI. XIV) and the horse-flies (Tabanidse, PI. X-XII), the larvae of all these families being largely aquatic in habit. The distinguishing characters of these groups are given in the keys which follow. THE LOCATION. The Illinois Kiver runs through one of the most re- markable and interesting valleys in this country. Al- though this valley is from one hundred to two hundred feet below the level of the surrounding country, it often reaches a width of ten or even twenty miles between the up- lands on either side, rivaling in extent the great trough of the Upper Mississippi, and containing alluvial deposits of great depth and of enormous extent. In striking con- trast with these evidences of vast erosion is the quiet river that lazily winds its way through a network of shallow weedy lakes and intricate sloughs—a paradise of aquatic life. It is evident that some greater stream once occupied this valley, by turns, no doubt, the outlet of Lake Michigan and the drainage channel of the great ice-sheet which is supposed to have thrown up the princi- pal drift ridge of central Illinois. Away from the low muddy flats along the river, with their lakes and sloughs, this valley is very generally flllod in to a considerably higher level with extensive deposits of comparatively pure sand, alternating with sheets of clay. A well at Havana seventy feet deep does not reach the bottom of this formation. The surface of these sandy Entomology of the Illinois River. 151 areas has partly been blown into undulating mounds and ridges, and occasionally presents expanses of barren sand. They are quite pecuhar in their fauna and flora, which are very different from those of the uplands. At Havana the width of the sand deposit between the river and the eastern upland is about fifteen miles. Just above Havana, Spoon River enters the Illinois from the west, and the sediment brought down by it has raised the bottom-lands below its mouth to a higher level, narrowing the river, and forcing it for a short dis- tance against the margin of the sandy plain. Above the mouth of Spoon Kiver the bottom lands are lower. Here wide stretches of shallow lake and marsh appear, with bottoms of soft black mud, and com- paratively narrow intervening wooded ridges, rarely sandy, except at certain exposed and wave-washed points or along the margins of the sand plain. THE SUBSTATIONS. In order to cover a variety of situations, a number of typical points or substations were selected for special study and periodical examination. Each of these was regularly searched from the shore into deep water; but as the forms here treated are essentially shallow-water forms, the characteristics of the marginal surroundings will receive special attention. Three of these stations w^ere located in Quiver Lake. This is a permanent arm of the river extending north- wards along the margin of the sand plain above Havana. The natural drainage of the sand escapes in large quan- tities along its eastern side, keeping the shore con- stantly saturated with cool percolating water, to a greater or less width, according to the level of the river. Beyond the head of the lake, this drainage forms a con- siderable stream. Quiver Creek, which empties into the broad shallow head of the lake through a muddy and weedy flat. Near the place where the clear waters of this stream cease to follow a definite channel, Station 152 Illinois State Laboratory of Natv/ral History. A is located. Not far below this issues the most ex- tensive flow of water from the muddy and sandy shore on the east side of the lake, maintaining there a consid- e: able variety of plants and animals left upon it in spring by the receding waters. This is Station B. The lake itself is moderately shallow, and is filled during low water in summer by a dense mat of Ceratophyllum, Anacharis, (Edogonium and other algae, with areas of Nelumbo, Nymphsea, Vallisneria, and Potamogeton. Through this mass of vegetation the clear water of the lake slowly moves towards the river, reinforced by the constant in- flow. Station C is located near the outlet of the lake, the shores here being near together and sheltered. The east bank is sandy, with a muddy coating over the part which is exposed at low water, while the west shore is of black mud grown over with willow trees and over- flowed in moderately high water. The water on both sides is thickly filled in summer with algae and other aquatic plants. Here during the first season our cabin- boat was located, giving us a greater opportunity to make observations at this station than at some of the others, to which fact is due a slight preponderance of data from this place. At the opening of the second season (1895) another point, Station L, was selected in Dogfish Lake, a branch of Quiver Lake on the west, also matted with vegetation but without flow of spring water. Near Station C, in the river itself, is Station E, a gently sloping muddy shore with but little vegetation; and in a weedier but more exposed position at the side of t]he broad expansion of the river known as Havana Lake, is Station D. In the narrower and more rapid part of the river below the city is Station H, with steeper shores, sandy and gravelly on the east and muddy on the west. Between Havana Lake and the margin of the sand plain at the east is a low bare island, separated from Entomology of the Illinois River. 153 the sand by a small grassy and weedy slough, spar.-jely grown with willows, through which the escaping spring water makes its way to the river. This peculiar slough is known as Station I. The three remaining stations are on the west side of the rivei', in lakes which have no noticeable inflow of percolating water: Phelps Lake, temporary in nature, now open water, now dry, communicating with the river only at rather high water, gives us Station F; Thomp- son's Lake, large and permanent, with almost constant river communication, mostly open water, affords us Station G; and Fkig Lake, a broad swampy expanse, widely' margined with club-rush (Scirpus), and with a line of open ponds, is studied as Station K. The location of the different stations and their char- acteristics at ordinary stages of water may be summa- rized as follows: A. Junction of Quiver Creek and Quiver Lake; shallow, mud and sand, grass and floating vegetation, variable. B. Wet springy shore of Quiver Lake; sand and mud, grass and coating of algte. C. Near foot of Quiver Lake; shores moderately slop- ing and sheltered, the eastern sand and mud, the west- ern mud, low, and wooded; water clear, with dense aquatic vegetation, having little or no current in ordinary stages. L. Dogfish Lake; arm of Quiver Lake, shallow, very gently sloping, mud, much floating vegetation, dead water. E. Riv^er near Station C, somewhat narrow but deep at middle; margin rather shallow, soft mud, little vege- tation, current slight, bank wooded. D. Exposed shore of Havana Lake (a broader part of river); moderately sloping, mud, considerable grass and aquatic vegetation, very little current, not wood- ed. H. River below Havana, narrow, east bank steep and sandy, a layer of mud over sand at lower levels, water 154 niinois State Labm^atory of Natural History. quickly deepening, considerable current, a little vegeta- tion, west bank of mud, low and wooded, steeply slop- ing, water almost without vegetation, decided current. I. Bed of the "Slough," with grass, rushes, and wil- lows,' and a very shallow stream of spring water when river is low. F. Phelps Lake; very shallow, often entirely dry mud, almost no vegetation, dead water, shores densely wooded. G. Thompson's Lake; exposed sandy shore, moder- ately sloping, grass and considerable aquatic vegeta- tion, bottom of lake muddy, dead water. K. Flag Lake; shallow, muddy, bordered with rushes, thick with floating vegetation. LIFE AT THE DIFFEKENT STATIONS. It is not too sweeping a statement to say that the full lists for each station of every species observed there during the year do not conspicuously differ. On the other hand, variations in relative abundance of the forms at each station and of the total life at each, with the presence or absence of some prominent species or group of species, impart an individuality to each station. These lists are unusually large, and I have at hand a much greater variety and number of species than a single locality will usually furnish. It will be best therefore to point out merely the leading differences, with special reference to the forms herein reported upon, leaving tbe full treatment of the subject until the material collected has been more fully studied. Evidently the main requisites of insect life are food and protection from enemies. An abundant growth of aquatic vegetation, therefore, supports a large number and variety of insects; those which find in it food and shelter, as do many Diptera and the case-flies, and those which prey upon the plant-feeding forms and upon each other while sheltered from vertebrate enemies. The Entomology of the Illinois River. 155 stations in Quiver, Doo'fish, and Thompson's lakes (A, B, C, L, and (i), and D, in the broader part of the river, are situations of the above description. Station A has no definite shore, but because of its shallowness exhibited many features of shore life, and varied greatly as the real mouth of the creek shifted up and down with the changes in the height of the river. Aquatic caterpillars found food here, and horse-fly and crane-fly larvae lived in the mud ; certain case-flies were more common because of the running water; soldier-fly larvae (Stratiomyiidae), well protected by their tough skin, found appropriate food; and predaceous bugs and beetle larvap also abounded. Surface-beetles (Gyi-inidae) were seen only now and then as single examples, and the larger surface-bugs or water-striders (Hydrobatidse) were never very numerous, though the smaller ones (Veiiidae) were often seen on the floating vegetation in immense numbers. Water-beetles, except the haliplids, were as a rule comparatively few, seeming to prefer grassy margins and sticks and logs. Chironomid larvae were abundant in the vegetation and the mud of the bottom. Top- rainnows were seen gliding about in the little clear spaces; leeches and mollusks were plentiful. At Station C the life was quite similar to the above, but the muddy west shore, often flooded and washed bare by the river, seemed inhospitable to shore-loving forms, beetles and dipterous larvae of this class being noticeably few, although small surface-bugs, stratiomyiid larvae and aquatic caterpillars were common in the abundant vegetable growth. The neuropteroid forms were here unusually abundant. The shore of Thompson's Lake at Station G, although exposed, was protected during the summer by a belt of algae and "moss" (Ceratophyllum etc.) and differed from the west side of C in its sandy shore, favorable to the development of dipterous larvae generally, horse-fly larvae and Stratiomyiidae often abounding in the rubbish 156 Illinois State Laboratory of Natural History. washed upoQ the shores at times of higher water. The definite grassy margins also favored water-beetle life, as was the case at Station D, which became very similar to the west shore of C as the river fell and the current nearly disappeared. Station B and the east shore of Station C, also belonging to this group, located on the east shore of Quiver Lake, introduce a new element—the belt of satu- rated sandy shore, uniform in temperature, coated with algal growth, and teeming with life, including vast num- bers of Asellus and Gammarus and an abundance of dipterous larvae. Leeches and spring-tails (Podura and Isotoma) were unusually common. At the upper station, B, the surfaces of bare reeking mud were inhabited by soldier-fly larvae of the genus Stratiomyia. The little hollows produced here by the tramping of cattle quickly filled with water, and in these hollows, as well as in similar protected depressions along margins everywhere, the mosquito larvae were noticeably abundant in due season, occurring also less commonly in open pockets in thick floating vegetation. Lack of a suitable food plant prevented the occurrence of aquatic caterpillars. Station E is like the west shore of C, except that a great reduction in the quantity of vegetable life cor- respondingly reduces the number of the more delicate shallow-water forms. There is a considerable amount of sticks and rubbish, and the harder-shelled water- beetles and water-bugs are therefore more numerous. Gyrinidae in small schools also appeared. The most aberrant station of all, Phelps Lake^ the bottom of which was almost as bare as a floor, was apparently equally destitute of life. Close search, however, revealed abundant Corisidae in the deeper water; Berosus and other small water-beetles in the shallows hiding under fallen leaves; Notonecta about a fallen branch in the water; and Heteroceridae swarming over the mud at the margin, in their tiny mole-like burrows. Entomology of the Illinois River* 157 FOOD RELATIONS OF THE GKOUP8 TREATED, The plant-feediDg aquatic caterpillars seemed to occur wherever their food plants grew. They are pretty well protected by their habits of coucealinent in cases or between leaves, the Hydrocampa apparently being kept in check largely by its hymenopterous parasite. The sand-wasps (BembecidoB) and the spider-wasps (Pompilidae) were frequently seen flying about. The former provision their nests with flies; the latter with spiders, which are themselves predaceous and abund- ant along wet shores. The effect of the sand-wasps on aquatic Diptera is presumably injurious, and that of the spider-wasps beneficial. The larvae of soldier-flies (Stratiomyiidse) herein treated seem to feed on minute plant life, Odontomyia as a rule taking that on aquatic vegetation, and Strat- iomyia that on wet muddy surfaces. Their tough skin efficiently protects these species from insect enemies, in both larval and pupal stages, and it is probable that like the Hydrocampa, they also are kept in check largely by their hymenopterous parasites (Smicra). Megill'i maculata, a coccinellid beetle, has been observed feeding on the eggs. I do not know of their being eaten by fishes. The larvae of horse-flies (TabanidaE;), except possibly the little Chrysops larvae, are active and rapacious. They apparently do not attack operculate univalves, but are known to eat those in which the opening of the shell is not protected by an operculum. I have not noticed, however, that they have any preference for these. They did not attack the blue earthworm {Sj^ar- ganophllus eiseni)—so abundant in their usual habitat; — when placed in confinement with them. Breeding-cn-e experiments lead me to think that their chief food is soft larvae. They are usually found in the light or sandy substance of wet shores. Although crane-flies (Tipulidae) oviposit freely in such places, their larvae 158 Illinois State Lahoratoinf of Natural History. never become excessively abundant, being probably a prominent element in the food of the horse-fly larvae. The latter, when washed out of their positions in the soft bottom of shallow waters or in the sand of the margins, float exposed upon the water and become an easy prey to fishes. A very efficient check on their in- crease is the hymenopterous egg parasite, which often destroys a large percentage of the eggs in an ^^^ mass. The aquatic crane-fly larvse (Tipulidae) feed as a rule on minute algae and the like, but one species observed is in all probability predaceous. Their probable relation to the horse-fly larvae has just been mentioned. I have not observed any hymenopterous parasites upon them. Predaceous Coleoptera and their larvse are often asso- ciated with them and other small dipterous larvae on wet shores. The food of the larvse treated under the name Leptidae, I am unable to discuss at present. A careful study of stomach contents is of course necessary before the food relations of the above forms can be definitely and fully described. METHODS. Collecting.—\\^^qqX^ in vegetation and on or in the bottom were taken by means of a dip-net—a net of about equal depth and width attached to a strong semi- circular ring firmly fixed to a long handle, the straight side of the ring being opposite the point of attachment. (See Plate XV.) For the larger and more active forms, a coarser net of minnow netting was used, and for smaller forms, one made of bobbinet proved most durable and sat- isfactory. To collect from the mud of the bottom, the wa- ter immediately over it was violently stirred and then swept with the net. The surface layer of mud, was also scooped up in the fine dip-net and then allowed to wash through, leaving the coarser contents in the net. In a similar way, insects on the bottom in deep water were secured by using a dredge, and washing its contents through Entomology of the Illinois River. 159 ^ series of net sieves. The aquatic vegetation, when free from mud, was violently washed in a large pan, many smaller forms being^ thus dislodged and coming to the surface. Insects occurring in open water were taken in drawing an ordinary towing-net. A whole world of minute insect life largely passes through these nets, which would clog up badly were they made of finer material, and a set of wire sieves was therefore used, the lowest one of very fine brass wire gauze, through which the mud or sand of the bottom and margins was sifted, the washing in the fine sieve be- ing diluted and examined ; or the bottom of the sieve was held slightly below the surface in a large vessel of clear water, and the contents gently stirred and closely examined. These fine washings from mud or thick vege- tation are often well worth saving in bulk. The minute life of weedy waters may be well collected by means of a Birge net, which is a small, deep, fine net, the opening guarded by a coarse wire gauze cone, its apex outward, which parts the vegetation as the net is drawn through it. The contents are removed by unscrewing a small cap from a short tube inserted at the narrow bottom of the net. Preserving.—Methods of preservation have \Qvy greatly improved of recent years, but much remains to be done before all kinds of material can be satisfactorily pre- served. The best results with most larvae of any size were obtained by heating them in water, not too rap- idly, to about 200° Fahr., and setting aside till cool. A small percentage of acetic acid will prevent the collapsing of very soft larvae. The principal trouble with this method arises from the expansion of the air within, but a slight inflation, especially in the crane-fly larvae (Tipulidae), is desirable, as it fills out the anal prominence and its soft appendages. This method is not suitable for pupae generally, nor for ephemerid nor perlid larvae with their flat gills. As a general preservative for material so pre- pared and for other large insects, we have depended 160 Illinois State Lahoratory of Natural Ilistoinf. chiefly upon 80% alcohol and water. Experiments with formaline indicate that it will satisfactorily preserve small and easily penetrated forms. Breeding.— For merely rearing insects, the best cage proved to be a wide-mouthed glass vessel varying in size according to the insects to be reared. The mouth may be covered when desirable by a piece of "Swiss"^ or cheese cloth held in place by a rubber band. For most work, cylindrical battery jars five or six inches in. diameter and about seven inches deep were very useful. In these, by attending to a few simple rules, one can- rear successfully almost any kind of aquatic insects, ex- cept such as require running water. Direct sunlight must not fall on the vessels, as it often overheats the- water; and hard water should not be used. We repro- duce so far as possible the natural surroundings where the insect occurred, using material from the place where it was found. The water need not be changed if no film appears upon its surface and healthy animal and plant life are present; but if a film develops and a foul odor becomes noticeable, actively decaying organic matter is- evidently present. This should be immediately removed, the water being frequently changed for a time. A very little water is often better than a large quantity, and too- much vegetation may be a disadvantage. A foothold above water for emerging imagos is desirable. Those forms which leave the water for pupation, such as the lar- vae of beetles, horse-flies and some Neuroptera, (Sialidse)^ are transferred when full grown to clean damp sand, cov- ered with light, fresh rubbish for shelter and moisture. Beetle larvae will often pupate under chips in such a place; and pupae in puparia may be placed on damp sand. Pupae will often be killed by suffocation if collect- ed in a bottle of water. Some isolated individuals should be reared to verify results. For keeping insects under continual observation and in natural conditions a square box, eight inches each- way, was used, two sides and the bottom of wood, the Entomology of the Illinois River. 161 other two sides of wire gauze. The upper rim is all of wood and supports a close fitting glass cover in a wooden frame, overhanging outwardly, so that the whole in- terior is visible through the glass. (See Plate XV.) These cages are so placed in the lake or river as to be about half full of water, thus maintaining the quality of water and the temperature natural to life occurring there. CHARACTERS USED IN THE KEY TO ORDERS. A typical insect larva or pupa has about a dozen usually well-marked divisions. The first is the head, the next three constitute the thorax, and the remainder the abdomen. In the stratiomyiids the thoracic segments are closely like those of the abdomen, but there is usually a noticeable difference. In most larvae the thoracic seg- ments are readily distinguished by the pair of jointed feet on each, and in the pupa or older nymph, by a con- spicuous backward extension from each side of the second thoracic segment, more or less covering a smaller pair of similar extensions on the third, these being the two pairs of wing-pads. The second pair is rudimentarj^ in the Diptera. The distinction between the larvae of groups A and B is not applicable to very young indi- viduals, but one will easily learn to know those of the first group, as they closely resemble the older stages, and the inter-resemblances are quite marked in each of its subdivisions, which are not numerous. The spring-tails are minute Avingless insects, many of which frequent wet shores, often hopping on the surface, but rarely discovered beneath it. The next three orders take up air from the water, usually by flat mem- branous gills upon the thorax, or the sides or the end of the abdomen, or within the end of the alimentary canal. The terminal setae are long antenna-like tails. The true bugs breathe air directly, those which swim in water coming to the surface for it. In group B the abdomen often bears beneath fleshy jointless prominences, used as feet and called false feet. —11 162 Illinois State Laboratory of Natural History. When gills are present they are usually filamentous or thread-like and borne on the sides or back. The spiracles or stigmata are the external openings of the internal air- tubes or tracheae. The separation of the pupae of this group has proven difficult. The dipterous pupa when it develops within the old larval skin usually gets air from spiracles at the- posterior end, but the free pupa, with legs and wing- pads visible, receives its principal air supply from the spiracles of the first thoracic segment (prothorax). These are usually well up on the dorsal surface, large or prominent, and very frequentlj'^ borne upon long antenna, like prolongations. On the other hand, the prothoracic spiracles in the remaining orders are on the sides at the hinder edge of the segment and never very conspicuous. The antennae of the pupa rest against the surface of the body, but may be known by their origin on the upper surface of the head. KEY TO THE ORDERS OF IMMATURE AQUATIC INSECTS. In this key extensive use has been made of the chapter on Insects by Dr. E. Schmidt-Schwedt in "Die Tier- und Pflanzenwelt des Stisswassers." The eggs can best be determined at present hj an examination of the various ways in which they are deposited, as illustrated by the figures which will be published in the course of this work. Nymphs (larvae and pupae) similar at all ages; wing- pads present except on very young nymphs, wanting in Thysanura; thoracic legs always present and functional; no abdominal legs. ...A. Larvae without wing-pads; pupae with wing-pads and with thoracic legs visible but not in use;* pupa sometimes concealed in a hardened footless larval skin B. * The pupae of Phryganeidse use their thoracic legs on leaving their cases for emer^ gence. Other exceptions to the characters given in these l^eys will probably appear as our knowledge increases. They can only approximate the truth at present, but may serve as a stepping-stone to something better in the future. Although they are dravi'n up for aquatic forms exclusively, it seems to me that the natural relationships of the different orders are here very unmistakably exhibited. Entomology of the Illinois River. 163^ A. Nymjphs. With biting mouth parts. Maxillae and mandibles retracted, apices only visible;^ minute semi-aquatic air breathers. (Spring-tails) Thysanura. Maxillae and mandibles prominent; water breathers^ usually with tracheal gills. Abdomen with terminal setae. Gills on thorax; setae usually two. (Stone-flies) Plecojotera. Gills on side of abdomen; setae usually three. (May-flies) Ephemerida. Abdomen without terminal setae, and with ter- minal flat gill-plates or internal gills. (Dragon-flies) Odonata. With jointed beak; air breathers. (True Bugs) Hemijptera^ B. LarvcB. With jointed thoracic legs. False legs wanting (except in a Philh3^drus, which has six pairs present) ; not living in a case. Filamentous gills wanting—present but not seg- mented in Berosus and Gyrinidae, segmented in Cnemidotus, in which the legs have but one claw; abdomen usually with terminal spiracles. ( Beet) es ) Coleoptera. Body with segmented filamentous gills; no spiracles at apex of abdomen; two claws on legs. Neu/roptera, A pair of false legs on last segment, each with one or two strong claws; usually in a tubular case. (Case-flies) Trichoptera. Five pairs of false legs, provided at their apex with numerous booklets; no spiracles at apex of abdomen. (Caterpillars) Lepidoptera. Without jointed thoracic legs. (Flies) Diptera. 164 Illinois State Laboratory oj Natural History. B. Pupae. Prothoracic spiracles small and lateral, or wanting. Appendages mostly free from each other. Pupae formed on land ; without gills. Antennal joints fewer than twelve. Coleoptera. Antennal joints numerous, more than twelve. Neuroptera, Pupae formed in the water in a case; with filament- ous gills. Trichoptera. Appendages, head, and thorax all united. Lepidoptera. Prothoracic spiracles dorsal, prominently developed, often borne at the end of antenna-like appendages; or the pupa is concealed in the hardened last larval skin. Diptera. LEPIDOPTERA. Two groups of this order may be considered here; those epecies which feed internally on aquatic plants and breathe by spiracles, probably making no use of the air contained in the water, although it seems that some can swim from one plant to another when occasion requires; and those which feed externally, being provided with means for appropriating oxygen from the air in the water, at least in the earlier stages, and therefore completely aquat- ic during a part of their life; living in cases or shelters usually covered exteriorly with green plant tissue, and fill- ing these cases with air during the pupal and sometimes also during the later larval period. To the former group belong certain Noctuidae, such as Arzama and Nonagria, and a pyralid (Pyrausta) herein treated ; to the latter group belong the members of the remarkable and inter- esting pyralid subfamily Hydrocampinae. Entomology oj the Illinois River. 165 CHARACTERS' USKO IX TABILA PING AQIJA I'lC LlJl'IOOPTEKA. The characters separating the NoctuidcS and P.yralidae in the following: key are those developed by recent at- tempts at a more natural classification of the Lepidop- tera, and are of wide application in separating; higher and lower forms. The distinctive features of Cataclysta are taken from Guenee. A piliferous tubercle is typically a darker and slightly elevated spot, bearing one or more stiff hairs or bristles. In a typical noctuid larva the first and uppermost of these, as seen on one of the middle segments of the abdomen, is on the middle fold of the segment, subdorsal in situation; the second is lower down, and on the posterior fold; the third is be- neath the first but above the spiracle; the fourth is pos- terior again; the fifth beneath the third and below the spiracle. In the pyralids the arrangement of the fourth and fifth is different, as shown in the key. KEYS TO AQUATIC LEPIDOPTERA. Larvae. Middle abdominal segments with the fourth and fifth piliferous tubercles approximate or united. (Pyralidae.) Ocelli five. Body with long respiratory filaments. [Fig. 1.] Paraponyx, Body without respiratory filaments. Elongate, moniliform. Gatachjsta, Rather thick at middle, slightly flattened, ends tapering. [Fig. 10.] Hydrocampa. Ocelli six. Pyrausta nelumUalis. Middle segments of abdomen with the fourth and fifth piliferous tubercles distant from each other. (Noc- tuidsB.) Ifonagria, Arzama. 166 lUinois State Laboratory of Natural History. PupcB. Seventh abdominal segment freely movable in male^ appendages sub-external, separated from internal cavity by well-developed partitions, and breaking^ away from it at time of emergeuce. Developed spiracles present on segments 2-4 of abdomen; its apex merely pointed. Spiracles of abdominal segments 2-4 about equal in diameter; ventral sheath surpassing sev- enth segment. [Fig. 6.] Paraponyx, Spiracles of second abdominal segment much smaller than the next two pairs. Ventral sheath long, nearly as long as abdo« men. f JHydrocampa iccittsalis, f Cataclysta, Ventral sheath short, not surpassing seventb abdominal segment. (Hydrocampa, fide Guenee.) Hydrocampa ohliteralis. Developed spiracles present on segments 2-7 of abdomen, its apex flattened, with mass of recurved chitinous filaments. Pyrausta nektmbialis. Fifth and sixth abdominal segments freely movable, seventh and remaining segments united in both sexes; appendages sub-internal, with only a slight separation from internal cavity, and not breaking away from it at time of emergence. Wonagria, Arzama. No Arzama or Nonagria were collected at Havana during the season, and further treatment of the Noctu- idsB is omitted at present. Family PYRALID^E. Nymph^ella. The wings of the moths are rather long and narrow, obtuse at apex, the primaries without transverse lines^ and more or less spotted or tessellated, the hind wings Entomology of the Illinois River. 167 whitish, faintly lined or dotted; ocelli wanting. The larva is unknown to me. N. maculalis Clem, [dispar Grote). These pretty moths were common about the small lakes in northern Illinois August 4, being taken at Sand and Fourth Lakes. They also occurred at Cedar Lake June 19 and August 11, and at Fox Lake June 22; and were taken in Urbana July 8 and 22 at electric lights. Paraponyx. The reasons for using this genus name are discussed under Hydrocampa. The moths are of a different facies from that genus, having narrower wings, and longstraight dark lines on the posterior pair. The larvae differ from all other known forms of the order in being provided with filamentous tracheal gills. Their habits are hence completely aquatic. It is stated that when the cocoon is formed the air vessels of the leaf to which it is attached are tapped, and the water in the cocoon is driven out and replaced by air. P. ohscuralis Gr. [Fig. 1-7.] The favorite food of the larva of P. ohscuralis is the leaf of Vallisneria spiralis^ and it has also been found upon Potamogeton nutans. The Vallisneria grows freely at Station A, trailing in the slow currents at the mouth of Quiver Creek, and here almost all our larvae were found. They feed at first exposed on the leaf, but later two or even three leaves are loosely webbed together face to face by each larva, between which it remains con- cealed while feeding. They are therefore hard to dis- cover unless their hiding places are broken up by seining or the like, when the larvae may be seen swim- ming about. In these retreats a fine but dense cocoon may be found spun by the larva, in which the pupal stage is passed. Full grown larvae and pupae were obtained from these situations in July. In August, how- 168 Illinois State Laboratory of Natural History. ever, large quantities of detaclied cases [i'^ig- 2] were found floating at the surface in the same vicinity—rough- looking oblong or triquetral cases about an inch long, formed by severing in a ragged way the connections of the larval retreat at either end. In these were pupsB and good-sized larv£e, and it must be the usual habit of this species to come to the surface in this manner before pupation, and live in a detached case, like other Hydrocampinse. In single instances, pieces of the leaves of Nelurabo and Potamogeton entered into the construction of these cases, and a side of one case was smoothly built up of Letnna trisalca, which was abundant there at that time. These cases were seen also in smaller numbers in September and October, at the same place. The ima- gos were quite generally distributed and abundant along the lake and river shores in June, July, and the first part of August—an earlier period than that of Hydrocam.pa ohliteralis, which was still common in Sep- tember, while the present species was not noticed after August 24. In the breeding-cages they emerged July 18 and 26 and August 1. They have also been taken in Ur- bana at electric light August 18 and 19. Young larvae, doubtless of the new brood, were seen on the Yallisneria July 23 in great abundance, and it is probable that the species hibernates in the larval stage, a medium- sized example having been taken at Station A De- cember 17. J£ntom.ology of the Illinois River. 1G9 The recorded occurrences of this species at Havana throughout the year may be tabulated as follows: RECORD OF PARAPONYX OB.SCUKALIS FOR THE YEAR/ Batei. 170 Illinois State Laboratory of Natural History. remaining gills arise from the posterior side, the branches leaving the stem at approximately equal distances on its basal portion, each branch being about as long as the portion of the stem beyond their point of union [Fig 4]. There are just one hundred of these filament- ous gills. The head, prothorax, and terminal segment are destitute of them, the mesothorax lacks one pair^ and the penultimate segment lacks all hut one pair; the remaining nine segments have the full number of ten to each segment, five on each side, which Muller* desig- nates as follows ; an anterior and a. posterior one form- ing the suprastigraal pair, a similar infrastigmal pair, and a pedal gill below these [Fig 5]. The number of branches of the different filaments may be expressed by a definite formula for each stage of larval life, notwith- standing the fact that a slight variation from it is very common, there being often one branch more or less in one or all the filaments of a series. By studying a number of individuals I endeavored to eliminate this variation, and the branches in a full grown larva may be tabulated accordingly as follows, the segments back of the head being numbered consecutively: *Zool. Jahrbuch. Ab. Syst., VI.. p. 626. t a., anterior: p., posterior; s., suprastigmal; i., infrastigmal; ped., pedal. Entomology of the Illinois River. 171 Siuce the uiiniber of branches is pioportionally nearly as above in all individuals examined, it will be sufficient in the following comparison of larvae of differ- ent ages, to state the formula for the first four abdom- inal segments, adding also Miiller's figures for P. stratio- tata, a European species. Species. 172 Illinois State Laboratory of Natural History. Thoracic segments very finely granulose, darkest at anterior margins. Thoracic suprastigmal gills rather near together, and at the same distance from the median line; posterior infrastigmal directly beneath the posterior suprastigmal, the anterior infrastigmal higher up, and close to the anterior mai'gin of the segment. On the abdominal segments [Fig. 5] the anterior suprastigmal IS nearer the median line than its mate, and it will be noted that the latter has the more branches, instead of the former, as on the thorax; the anterior infrastigmal is lower than its mate, and more nearly in line with the anterior suprastigmal as compared with the arrange- ment upon the thorax; the pedal gill is on the latero- ventral fleshy prominence lust above the base of the leg, and is wanting on the mesothorax. The ninth ab- dominal segment bears only a single pair of gills aL the posterior margin, which appear to be the posterior infrastigmals. The first piliferous tubercle is imme- diately in front of the anterior suprastigmal on the tho- rax, and just within it on the abdomen; the second and third are in their usual places, the former not far beneath the posterior suprastigmal. A pair of short setae, one above the other, placed between the infrastigmals, appear to represent the fourth and fifth tubercles, and the sixth or seventh is immediately in front of the pedal gill, re- placing it on the mesothorax, and bisetose on the pro- thorax. Spiracles minute, oval, anterior ones slightly elevated, those at middle of body in a more or less dis- tinct yellow-brown spot, often enclosed by a darker ring. True legs rather stout ; sutures brownish ; claws black- ish, base paler, a small basal lobe; tibiae with a subapi- cal verticil of hairs. False legs short, five pairs as usual ^ first four with an oval entire circlet of hooks, the hooks with longer bases alternating with one or two shorter ones, the whole surrounded by a variably distinct fine blackish ring; circlet of hooks on posterior pair diminish- ing back of middle on each side and coming to an end at the middle third [Fig. 3]. Entomology of the Illinois River, 17B Papa [Fig. 6, 7].—Length 9-11 mm., greatest diameter 2.5 mm.; spiracle-bearing segments broadest, slightly swollen dorsally, body with sides gently curving, and tapering gradually towards each end ; rather soft-bodied, pale 3'ellowish white, eyes darker; surface smooth, subo- paque, and nearly naked. Head small, with two small de- hiscent spike-like porrect seta? on the vertex. Spiracles of segments 2-4 of the abdomen large and conspicuous, about equal in size, transverse diameter slightly greater than the longitudinal, borne on rounded tubercles. Ven- tral sheath reaching a little beyond end of seventh seg- ment. Apex of abdomen subacute; ninth segment beneath with a faint elevated line at middle, and a small eleva- tion each side. Anterior margin of ninth segment con- spicuously elevated into a broad transverse ridge, bear- ing a row of seven sharp brownish or blackish short longitudinal carinse; ninth segment beneath with a Y- shaped impression [Fig. 7]. P. alhalis Rob. The larva of this species is doubtless very close to that of obscuralis, as the difference between the imagos is slight, though constant. One of the moths was captured Sept. 11 at Station C. They have also occurred in our collections from the small lake region of northern Illi- nois and southern Wisconsin; at Cedar Lake June 19; abundant at Sand and Fourth Lakes Aug. 3 and 9; Aug. 30 and Sept. 3 at Lake Geneva, Wis.; and at Dela- van Lake, Wis., Sept. 5 and 6. Aug. 4 it was taken in Urbana at electric light. P. allionealis Walk, {plenilinealis Gr.). This is closely related to the European P. stratiotata, whose larva has long been well known, and which differs from that of P. ohscuralis in the lesser development of the respiratory filaments, as was shown in treating the latter species. Moths of allionealis occurred at electric lights in Urbana May 29, June 7 and 29. and July 6. 7, 23, 27, appearing in numbers on the evening of June 29, 174 Illinois State Laboratory of Natural History. Chrysendeton. The moths of this genus and of Cataclysta are of similar appearance, having rather narrow wings, the hind wings with a series of small patches of metallic scales surrounded by a black background, near the posterior margin. Ocelli are wanting in Cataclysta, present in Chrysendeton. The immature stages are in all probability much alike, and the larvae aquatic. €ataclysta has not been taken by us in Illinois. C. cloMdialis AValk. {medicinaUs Gr.). This graceful moth appeared at the electric light in Champaign June 21. Grote's types were from Illinois. Hydrocampa. The moths of the three genera Hydrocampa, Oligostigma, and Paraponyx show but slight structural differences; the latter has however been usually maintained as distinct because of its remarkable larval structure, the larva of Oligostigma being hitherto unknown. In Smith's Check List, on the other hand, Prof. Fernald, doubtless appreciating the undesirability of separating these genera by means of the immature stages, at least so long as our knowledge of them remained so glaringly incomplete, has placed all the North American species oi these three genera under Hydrocampa, together with some which had previously been included under Homo- physa, among them H. oMiteralis. Our studies show that two very distinct types of larvae are thus included: one represented by the species previously assigned to Homophysa and Hydrocampa; the other, by those assigned to Oligostigma and Paraponyx. The general appear- ance of the adults would seem to confirm this group- ing. There seems to be a clear generic distinction pres- ent in the immature stages,—not only in the develop- ment of respiratory filaments, but also in the structure of the posterior feet of the larva and the spiracles of the pupa,—though perhaps not manifest in the imago. Entomology of the Illinois River. 175 I have therefore restored Paraponyx, including under it, ^s did Lederer,* the species of Oligostigma. Packard has figuredf the transformations of what he supposed to be II. Icckisal'os— a typical Hydrocampa. We have bred all stages of H. ohllteralls, which belongs to the group of Hydrocampas once classed with Ho- mophysa. There is no trace in our larva of the posterior pits fi<;' m1 by him in iccinsalis; and while the pupse seem to agree in relative size of spiracles, the ven- tral sheath in his figure is made very long, as de- scribed by Guenee for Cataclysta, this character seeming to be correlated with the length of the wings in the adult. The sketchy nature of the figures makes further comparisons uncertain. The wings of the Hydrocampa moths are broad or moderately so, the hind wings crossed by a pair of wavy lines near middle. Jff. gyralis Hulst. A single example was taken flying about the cabin- boat July 19. The species had been previously taken by us at electric lights near the University June 17 and Aug. 24, 1886. The three examples thus collected ara all males, as were also Mr. Hulst's types. We have in our collection also two undetermined Hydrocampas, both females, which, although differing greatly in general appearance from the preceding, and somewhat larger, are in all probability the other sex of this species, as the pattern is essentially the same, and the hind wings scarcely differ at all. The white lines of the fore wings, however, so sharply distinct in the males, are here obscured by a tawny yellowish suffusion, and the dark shades are indefinitely outlined. These also are from electric light collections made near the University in 1886, on May 28 and Aug. 23 of that year. *Wien. Ent. Monatschr., VII., p. 452. + Am. Kat. 1884, p. 824. 176 Illinois State Lahoratory of Nat/wral History. H. ekthlipsis Gr. Taken by us but once, July 5, at an electric light in Champaign. H. icciusalis Walk, [genumalis IjqA., formosalis Clem.). Although the moth of this species is the most common hydi'ocanipid about the University, it was not seen at all at Havana, The probable differences between the immature forms of this and the next species have al- ready been mentioned. We have found the imago at electric light in Urbana May 19, 31; June 3, 15, 26; July 5, 6, 7, 20, 21, 28 T Aug. 2, 17, 24; and at Lake Geneva, Wis., on Sept. 3. S. oUiteralis Walk, {proprialis Fern.). The favorite home of this species is among the float- ing leaves of Potamogeton nutans^ which often thickly cover the surface of quiet water in large patches. The amber-colored eggs [Fig. 8, 9] were first noted June 1, and are laid in a long baud just within the margin, on the lower surface, of some broad floating leaf, usually that of P. nutans. They are closely placed in a single layer, in rows running parallel to the margin, the band being about 3 mm. wide and including usually five or six rows of eggs, the members of each row alternating with those of the adjacent rows honeycomb-fashion. Their long axes point to the margin of the leaf, and each egg slightly overlaps those adjoining it on the inner side, showing that the moth probably rests at the edge of the leaf above and extends her ovipositor beneath it. The band is usually an inch or two long. One leaf of P. nutamjS in our collection, over three inches long, is en- tirely margined with eggs, except a short interruption at the side and another at the extreme base of the leaf. These egg bands were common in July, and a few were seen in August. Larvae were obtained from them in the breeding-cage July 13. The next day they had cut out minute oval disks from the leaf, and webbed these to its lower sur- Entomology of the Illinois River. 177 face, secreting themselves in the retreat thus formed, and feeding upon the substance of the leaf. When a little older, the larva cuts loose the portion of leaf surface to which it has attached its shelter, and is thereafter found traveling about like a case-worm in a lens-shaped case, formed of two ii-regularlj oval convex pieces of green leaf attached at their sides and open at the ends. In the larger cases made by older larvae the posterior end is narrowed, giving the case an ovate shape, or is even provided with a projecting median lobe, like the neck of a bottle, in which rests the pos- terior end of the larva [Fig. 11]. A case found by us in September had one side made of the fronds of Lemna trimdca^ and another was entirely formed of these little fronds. The pupal case is similar to that of the larva, but is smaller, oval, and more convex, and the edges are apparently strongly webbed together throughout, though the anterior end is easily parted, revealing an oval cavity with closely woven silken walls. When quite young, the larva is submerged and water-breathing, but soon fills its case with air and breathes it directly. In this respect it differs from Parapon^^x, which remains submerged throughout its larval life. The larvae [Fig. 10] were common in July and mostly reached the pupal stage during the first part of August, those seen towards the last of this month being fewer in number and mostly full grown. They continued to appear, however, and still occurred in the collections made in October. Tlie presence of more young larvae than usual was noted Sept. 20, and a young example taken in Flag Lake Mar. 23 seems to be of this species. A few imagos were seen at the time the first eggs were collected, and they continued to increase in numbers, becoming most abundant in August and September, when they were quite common on plants over water and often settled abundantly upon the sides of our row- boats. A few were seen in October. Thev emerged in -12 178 niinois State Laboratory of Natural History. the breeding-cage Aug. 8-10, having pasBed through the pupal stage very quickly. Examples of the moth have- also been taken in southern Illinois, at Cobden, June 12, and one appeared at the electric light in Urbana Sept. 4. RECORD OF HYDROCAMPA OBL1TERALI8 FOR THE YEAR.* Dates. Entomology of the Illinois River, 17^ where the largest numbers of the Hydrocampa occurred, flitting about in an investigating way over the floating leaves, often disappearing in the water beneath the^ edge of one and soon reappearing at some other point of its margin and flying to another. This Cryptus has, in fact, been previously bred in Florida, March 26, from a Hydrocampa on water lily, supposed to he olUtei'alis.*' Egg [Fig. 8, 9].—Length 6 mm., width 4 mm. Oval, amber-colored, flattened, one surface broadly gummed to the leaf, the other finely longitudinally wrinkled, a longi- tudinal elevated ridge at middle. Described from examples laid in a band on the under side of a floating leaf of Potamoyeton nvfan,?^ near and parallel to its margin. Larva^ first stage.—This does not differ markedly in sur_ face, structures, or color, from the mature larva; the setae are more conspicuous, especially posteriorly, and the ocelli are closely approximate, the lower three in a solid oblong black dash, with the other two just above. Larva, mature [Fig. 10—12].—Length 13-14 mm., width 2 ram. Fusiform subcylindrical, slightly depressed, broad- est at middle, uniform dirty whitish. Surface subopaque^ microscopically granulate or scabrous, more noticeably so on the anterior part of the thorax. Head rather small, light brownish yellow; Y-mark narrowly darker, bordered each side with whitish; a lateral brown stripe from base of head nearly to ocelli; labrum deeply and rather acutely emarginate at middle; mandibles sharply toothed ; sutures beneath more or less darker; first antennal joint truncate-conic, whitish, second slender, very pale yellowish, tipped with a seta and three minute articles, the middle one of which is deeply cleft; ocelli five, four in a slight curve just back of base of antenna, the middle ones contiguous, their pigment spots large and confluent, the upper one smaller and more isolated, behind it and similar to it, the fifth one. *"Bre